MaplePrimes - answers and comments on Question, 5) Look at the below code. It takes 7 seconds

The code that you've written does not execute in parallel

DuncanA — Thu, 26 Aug 2010 14:06:32 Z

The code that you've written does not execute in parallel.

According to ?Threads[Task][Start], "The Start function creates a new Task and then waits for that Task to complete before returning the value returned by the Task."

Each call to Start, for example,

Threads:-Task:-Start(X1);

will run to completion before the next Task starts.

So, in your code

Threads:-Task:-Start(X1);
Threads:-Task:-Start(X2);
Threads:-Task:-Start(X1);
Threads:-Task:-Start(X2);

the tasks will run sequentially, not in parallel. The ?Threads[Task][Start] help page gives an example of how to use Start and Continue to execute tasks in parallel.

---

Duncan

Sockets does not run in parallel

dohashi — Thu, 26 Aug 2010 22:23:31 Z

It is worth pointing out that the most of Maple's packages, including the Sockets package, have not been verified as thread safe, therefore they may not work correctly in parallel. Because we have not verified Sockets as thread safe and because it is implemented as an external library, Maple linearizes calls into the sockets external library. Therefore code like this will not show a significant speed up in parallel.

You can see my post on the limitation of parallelism in Maple for more information.

Darin

-- Kernel Developer Maplesoft

why do you use Maple for such?

Axel Vogt — Thu, 26 Aug 2010 23:25:55 Z

With Excel there is 'nothing' to code (just some configuring a sheet for such a query), see its help.

Exporting or processing data is a different thing, in any system I would guess.

And most of the time is spent on retrieving data from the web (you have to look up some Excel sites, but I guess if you feed various sheets in a book they are done non-exclusively).

query_example.zip

Edited: or better and more general one may wish to consider downloader programs like the old 'wget' (have not checked for Google) or similar writing to files, to be processed by what ever.

suggestion

Joe Riel — Fri, 27 Aug 2010 03:52:39 Z

The bottleneck is not in Maple but rather in the access to the external site. Wrap your call in CodeTools:-Usage:

CodeTools:-Usage(X1()):
memory used=275.27KiB, alloc change=0 bytes, cpu time=0.00s, real time=0.59s

The cpu time (0.00) is the Maple kernel time.

While it won't help the time (it's hard to improve on 0), you can reduce the memory usage by using ?StringTools[StringBuffer] to catenate the lines:

X1 := proc ()
local str, sid, b, buf;
uses Sockets;
    buf := StringTools:-StringBuffer();
    sid := Open("google.com", 80);
    Write(sid, cat("GET /finance/historical?q=INDEXSP:.INX&histperiod=daily HTTP/1.0 \n\n"));
    do
        b := Read(sid);
        if b = false then break; end if;
        buf:-append(b);
    end do;
    Close(sid);
    buf:-value();
end proc:

CodeTools:-Usage(X1()):
memory used=38.13KiB, alloc change=0 bytes, cpu time=0.00s, real time=0.56s

@DuncanA ok thank you for that

alex_01 — Thu, 26 Aug 2010 15:08:20 Z

@DuncanA ok thank you for that explanation.

Could you please explain to be how I can make the above
code work in parallel ie can you provide me with a modification
of the above code 5) so it can work in parallel so I can see the speed improvment
of calling google finance in parallel instead of sequentially...

If it is not to much to ask for could you also provide some modified code for 4)
so I can see the impovement in speed by calling X2+X3 in parallel compared to
calling X1. Since X2+X3 is called in parralel it should be much faster than X1.....

Importing Quotes and Counting 7s in parallel

DuncanA — Thu, 26 Aug 2010 16:59:07 Z

@alex_01 Here are two worksheets that perform the tasks in parallel. Both use Threads not Tasks. I'm using Maple 12 which doesn't support Tasks.

Counting_7s_with_Thr.mw

Importing_Quotes_wit.mw